基于CNN和XgBoost的香蕉成熟度判别

Banana ripeness determination based on CNN and XgBoost

目的:提高香蕉成熟度的判别准确率。方法:基于卷积神经网络和极限梯度提升算法建立香蕉成熟度的判别方法。先通过卷积神经网络提取香蕉图像特征,并采用全连接层网络和线性判别分析方法精简香蕉图像特征;通过贝叶斯优化算法优化极限梯度提升算法超参数;将简化后的香蕉图像特征输入极限梯度提升算法,通过极限梯度提升算法对香蕉成熟度进行判别。结果:所提方法对香蕉成熟度的判别准确度为91.25%;与已有方法相比,所提方法对小数据量香蕉的成熟度判别准确率明显提高。结论:该方法可实现被测香蕉成熟度的准确判别,有助于仓库经理、出口商实时监测香蕉的成熟度状况。

Objective:Improve the identification accuracy of banana ripeness.Methods:A novel method was established to identify banana ripeness based on CNN and XgBoost.Firstly,convolutional neural network was used to extract banana image features,and full-connected layer network and linear discriminant analysis were used to simplify banana image features.Then,the hyperparameters of the limit gradient lifting algorithm were optimized by Bayesian optimization algorithm.Finally,the simplified banana image features were input into the limit gradient lifting algorithm,and the banana ripeness was judged by the limit gradient lifting algorithm.Results:The identification accuracy of the method for banana ripeness was 91.25%.Compared with the existing methods,the proposed method was more accurate to distinguish the ripeness of bananas with small data volume.Conclusion:The proposed method can realize the accurate identification of banana ripeness,which is helpful for warehouse managers and exporters to monitor banana ripeness in real time.